OHDSI Home | Forums | Wiki | Github

Why is UA, NSTEMI, and STEMI not hierarchically tied to Acute Coronary Syndrome

(Mui Van Zandt) #1

I have a question from my Chinese colleagues. They have 4 terms and according to their text books and clinical experience, the four terms should be related in a hierarchical form. The terms are: Acute Coronary Syndrome, UA (unstable angina), NSTEMI (Non-ST-elevation myocardial infarction), and STEMI (ST-Elevation Myocardial Infarction).

Shouldn’t UA, NSTEMI, and STEMI be children of Acute Coronary Syndrome (Concept ID - 4215140).

Thank you.

(Eduard Korchmar) #2

Hello! I’m Eduard from Vocabulary team.

There is definitely a problem with this branch of diagnoses. Acute coronary syndrome seems misplaced, as for Unstable angina — not sure if it belongs with acute disorders. Although it clinically may lead to acute infarctions or other acute disorders it itself is not a subtype of one.

All of these concepts are supplied by SNOMED CT. Entire Vocabulary is implemented in OMOP CDM ‘as is’, although we often encounter mistakes inside of it. Since SNOMED is updated by source following ever-changing guidlines regularly so we consider trying to implement and support our own fixes impossible at the time.

Instead at the time we collect reports and work with IHTDSO (organization maintaining SNOMED CT) to establish feedback process. We can’t say with confidence how soon changes can be applied. Could you please share your use-case so we could suggest a work-around in the nearest time?

(Christian Reich) #3


As @Eduard_Korchmar says - we just import what SNOMED has. And they have a funny hierarchy:

  • The EKG findings have their own hierarchy (should be Measurements, can you fix, @Eduard_Korchmar and @Dymshyts?)
  • STEMI and NSTEMI are children of myocardial infarction
  • ACS has no social life.

We should discuss this with SNOMED.

(Dmytry Dymshyts) #4

It is not related to this discussion directly, but it is an interesting question.
My impression that MEASUREMENT is something like a Variable - Value pair.
But SNOMED EKG findings are actual findgings (only Values), such as
ECG: junctional tachycardia

EKG: atrial ectopics
that’s why they belong to Condition domain.

(Alexander Davydov) #5

Dmytry, not only values, but standardized tests + values in one concept. Fits to Measurement domain definition. We discussed it here before.

(Christian Reich) #6


(Eduard Korchmar) #7

There is a huge branch of a parent ’ 40480457 Evaluation finding’ including EKG findings and many similar results of measurement/evaluation procedures.

We intend to research this branch and make all concepts from it Measurements (given they can be Measurements), but it might take some time. We would rather include this task it in the next release.

In the meanwhile, as a temporary solution, we could make just EKG finding concepts Measurements at this time, if it really is an urgently needed fix.

(Manlik Kwong) #8

Hi - if you are trying to utilize EKG data to identify ACS, UA, NSTEMI, STEMI and using commercial electrocardiograph machines, you can use measurements and/or the computerized interpretation statements mapped to measurements, observations, and conditions.

Using measurements if you want to explore your own EKG criteria. Using computerized interpretation statements if you want to rely on what clinicians are seeing at the point of care.

We have OMOP mapping for the top commercial measurements and interpretation programs used in the US I can share with you if this would be helpful.

  • MK

(Christian Reich) #9

No need. Just change the “peaks” to create the proper solution.

(Christian Reich) #10

Yes, please. Would you be so kind?

(Manlik Kwong) #11

What is the best way to share these mappings? They are in EXCEL spreadsheets - one for each manufacturer measurements and interpretation statement sets. There are some measurements that are manufacturer specific and don’t have a LOINC mapping and have implemented as a custom vocabulary/concept locally (Philips Medical).

  • MK

(Christian Reich) #12

Can you upload here?

(Manlik Kwong) #13

There are 3 Philips Medical criteria - 09 is the measurements program, 0A, 0B, 0C are the three variants of the computerized 12-lead interpretation programs (attached).

GE Healthcare has the GE12SL interpretation program (attached).

The GLASGOW interpretation program is found in 12-lead defibrillators.

We are actively using the Philips ECG mapping at Tufts Medical Center. I hope others will find these to be useful and appreciate any feedback.Philips0AMapToOMOP_draft1.xlsx (132.9 KB)
PH100CMapToOMOP_draft1.xlsx (185.3 KB)
PH100BMapToOMOP_draft1.xlsx (145.3 KB)
ge12slMapToOMOP_draft1.xlsx (187.7 KB)
GlasgowMapToOMOP_draft1.xlsx (142.6 KB)